Near IR photometry of the old open clusters Berkeley 17 and Berkeley 18

Transcription

1 Astron. Astrophys. 343, (1999) Near IR photometry of the old open clusters Berkeley 17 and Berkeley 18 Probing the age of the Galactic Disc Giovanni Carraro 1, Antonella Vallenari 2,Léo Girardi 3, and Andrea Richichi 4 ASTRONOMY AND ASTROPHYSICS 1 Dipartimento di Astronomia, Universitá di Padova, Vicolo dell Osservatorio 5, I Padova, Italy (carraro@pd.astro.it) 2 Osservatorio Astronomico di Padova, Vicolo dell Osservatorio 5, I Padova, Italy (vallenari@pd.astro.it) 3 Max-Planck-Institut für Astrophysik, Karl-Schwarzschild-Strasse 1 D Garching bei München, Germany (leo@mpa-garching.mpg.de) 4 Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, I Firenze, Italy (richichi@arcetri.astro.it) Received 6 November 1998 / Accepted 22 December 1998 Abstract. We report on near IR (J and K bands) observations of two 8 8(arcmin) 2 regions centered on the old open clusters Berkeley 17 and Berkeley 18, for which only optical photometry (in B,V and I bands) exists. J and K photometry allows us to obtain an independent estimate of cluster metallicity by means of the relationship between the spectroscopic metallicity and the Red Giant Branch (RGB) slope calibrated by Tiede et al. (1997). From the analysis of the colour magnitude diagram (CMD) and luminosity function (LF), Berkeley 17 turns out to have a metal content [F e/h] It is 9 Gyr old, suffers from a reddening E(B V )=0.58 mag and has an heliocentric distance of 2.5 kpc. Berkeley 17 comes out to be substantially younger than in previous work (age 13 Gyr). On the other hand Berkeley 18 is found to have solar metal abundance, and to be younger than Berkeley 17, with an age of about 4 Gyr. While we confirm Kaluzny (1997) reddening estimate, we significantly revise the distance of the cluster, which lies 4.5 kpc from the Sun. These results on two open clusters believed to be between the oldest put constraints on the age and the evolution of the Galactic Disc. The absence of clusters older than 8 9 Gyr suggests the possibility that the Galaxy underwent a star formation minimum between 13 and 10 Gyr ago. Key words: Galaxy: evolution Galaxy: open clusters and associations: individual: Ber 17 Galaxy: open clusters and associations: individual: Ber Introduction Whether the Galaxy formed by an homogeneous monolithic collapse alá Eggen et al. (1962), or by a series of merging and/or capture events alá Searle & Zinn (1978) is still an open question. The traditional approach to investigate this fundamental issue is the comparative study of different stellar populations long ago Send offprint requests to: Giovanni Carraro (carraro@pd.astro.it) Based on observations taken at TIRGO. recognized to exist inside the Milky Way (halo, disc - thin and thick - and bulge). In particular a decisive step is to understand the formation and evolution of the Galactic Disc, and its relation with the Halo and the Bulge. Open clusters, especially the oldest ones, are the best candidates to trace the early stages of the disc development, the chemical and dynamical evolution of the disc, and the possible existence of a hiatus in the star formation history of the Milky Way between the formation of the halo/thick disc and the thin disc. Janes & Phelps (1994) discussed in details some of these interesting questions, suggesting that: the age distribution of the open clusters overlaps that of globular clusters suggesting that the disc started to form before the end of the halo formation; captures events were responsible for the formation of the disc radial metallicity gradient; the overdensity of open clusters in the age range between 5 7 Gyr is probably due to a burst of star formation induced by a merging episode. The first point is of particular interest. The conclusion of the authors completely relies on the existence of a cluster, Berkeley 17, whose age is estimated around 12 Gyr (Phelps 1997). The bulk of globular clusters has ages around 13 Gyr (Gratton et al. 1997) with a tail down to 11 Gyr. This opens the possibility that there has been no hiatus at all between the formation of the halo/thick disc and the thin disc. In this paper we would like to address the issue of the age of Berkeley 17, analyzing IR photometry in J and K bands. We also discuss similar data for Berkeley 18, another old open cluster. This is part of a project to study northern open clusters in the IR. We report elsewhere (Vallenari et al. 1998) on two very young open clusters, Berkeley 86 and NGC The age of Berkeley 17 has been discussed by Carraro et al. (1999) using published BV I photometry. From this study it emerges that Berkeley 17 is about 9 Gyr old, and the oldest open cluster, NGC 6791, is 3 4 Gyr younger than the bulk of globular

2 826 G. Carraro et al.: Near IR photometry of the old open clusters Ber 17 and Ber 18 Table 1. Basic parameters of the studied clusters. Cluster α δ l b Diameter hh mm o o o ( ) Berkeley : Berkeley : Table 2. Observation log book Cluster α δ Date Exposure Times (sec) (2000) (2000) J K Fig. 1. A mosaic of the four regions observed in the field of Berkeley 17. Be Oct, 27, Be Oct,25, Fig. 2. A mosaic of the four regions observed in the field of Berkeley 18. clusters, suggesting the possibility that a real star formation minimum occurred between the halo and disc settling. Berkeley 17 and Berkeley 18 were discovered by Setteducati & Waever (1962) in their search for unstudied open clusters. Both the clusters were never observed in the IR, whereas they have been studied in the optical (B, V and I). Berkeley 17 was studied by Kaluzny (1994) and Phelps(1997), whereas Berkeley 18 was studied by Kaluzny (1997). Their basic properties are summarized in Table 1. This paper is organized as follows: Sect. 2 presents the observation and data reductions; Sect. 3 discusses CMDs of the clusters under investigation, whereas Sects. 4 and 5 deal with the derivation of the clusters fundamental parameters. Finally Sect. 6 gives some concluding remarks. 2. Observations and data reduction J (1.2 µm) and K (2.2 µm) photometry of the two clusters was obtained at the 1.5m Gornergrat Infrared Telescope (TIRGO) equipped with Arcetri Near Infrared Camera (ARNICA) in October ARNICA is based on a NICMOS pixels array (gain=20 e /ADU, read-out noise=50 e, angular scale =1 /pixel, and 4 2 field of view). Through each filter 4 partially overlapping images of each cluster were obtained, covering a total field of view of about 8 8 2, in short exposures to avoid sky saturation. The two fields are located close to the center of Berkeley 17 and Berkeley 18. The log-book of the observations is presented in Table 2 where the centers of the observed fields and the total exposure times are given. The nights were photometric with a seeing of Fig. 1 and Fig. 2 presents the final mosaics of the 4 frames for both clusters in K passband. The data were reduced subtracting from each image a linear combination of the corresponding skies and dividing the results by the flat-field. We make use of the Arnica package (Hunt et al. 1994) in IRAF and Daophot II. The conversion of the instrumental magnitude j and k to the standard J, K was made using stellar fields of standard stars taken from Hunt et al. (1998) list. About 10 standard stars per night have been used. The relations in usage per 1 sec exposure time are: J = j (1) K = k (2) with standard deviation of the zero points of 0.03 mag for the J and 0.04 for the K magnitude. This error is only due to the linear interpolation of the standard stars. The calibration uncertainty is dominated by the error due to the correction from aperture photometry to PSF fitting magnitude. The standard stars used for the calibration do not cover the entire colour range of the data, because of the lack of stars redder than (J K) 0.8. From our data, no colour term is found for K mag, whereas we cannot exclude it for the J magnitude. Taking all into account, we estimate that the total error on the calibration is about 0.1 mag both in J and K pass-bands. 3. The color-magnitude diagrams K (J K) relationship for Berkeley 17 and Berkeley 18 are shown in Fig. 3 and Fig. 4, respectively Berkeley 17 The CMD of Berkeley 17 shows all the features of a very old open cluster, namely a scarcely populated extended Red Giant

3 G. Carraro et al.: Near IR photometry of the old open clusters Ber 17 and Ber Phelps (1997). The brighter part of our CMD is somewhat more populated than Phelps (1997) study, because of the larger field coverage (8 8 compared to ). As a consequence, the RGB is better defined in our study. The widening of the MS is due to many causes (binaries, differential reddening and so forth), the predominant one being the photometric errors, which are 0.06, 0.12 and 0.17 at K = 16, 17 and 18, respectively. The magnitude difference K between the TO and the clump turns out to be about 3.2 mag Berkeley 18 Fig. 3. The CMD of Berkeley 17. As for Berkeley 17, Berkeley 18 shows to be an old cluster. It appears to be less contaminated in field foreground/background stars, the MS turning off at K and (J K) The RGB shows some gaps, whereas the clump of He-burners (about 10 stars) is located at K and (J K) The magnitude difference K turns out to be about 2.7, suggesting that Berkeley 18 is younger than Berkeley 17. We notice that the CMD shows the presence of a group of stars above the clump, which does not lie along the RGB. These are probably field stars, which are also present in the photometric BV I survey of Kaluzny (1997). Also in the case of Berkeley 18 we ascribe the MS widening to binary and field stars contamination, possible differential reddening and photometric errors. These latter are 0.08, 0.10, 0.18 at K = 15, 17 and 18, respectively. 4. Berkeley 17: cluster parameters 4.1. Metallicity Fig. 4. The CMD of Berkeley 18 Branch (RGB) and a clump of He-burning stars. Some field stars contamination from the galactic disc is visible in the main sequence (MS), which extends almost vertically well above the cluster Termination Point (TO) up to K =11. The cluster TO is most probably located at K 14.5 and (J K) 0.50, whereas the clump of He-burning stars is visible at K and (J K) It is composed by 4 5 stars, which are the counterpart of the group of stars located at V 14.5, (V I) 1.7 in the BV I photometry of Friel et al. (1995) measured Berkeley 17 metal abundance using moderate resolution spectroscopy of 12 stars. They obtained [F e/h] = 0.29±0.13. It is possible to derive the cluster abundance photometrically in the IR using the calibration of Tiede et al. (1997). This relation correlates the abundance [F e/h] with the slope (J K)/ K of rhe RGB. It reads: [F e/h] = 2.98(±0.70) 23.84(±6.83) (RGBslope)(3) This relation has been proved to hold for both metal rich globular clusters and open clusters. For the latter the RGB slope tends to present less negative values at decreasing age. We applied this relation to Berkeley 17, for which the RGB slope is (J K)/ K = 0.11 ± This results in [F e/h] 0.35, consistent with spectroscopic estimates Age, distance and reddening Berkeley 17 was studied in the optical by Kaluzny (1994) and Phelps (1997). Kaluzny(1994) compared Berkeley 17 with NGC 6791 and, assuming that the former is metal poorer than the latter, concluded that the two clusters are probably coeval (around 9 Gyr old). Moreover he suggested a distance modulus (m M) > 15.0 mag and a reddening E(V I) > 0.70 mag. On the other

4 828 G. Carraro et al.: Near IR photometry of the old open clusters Ber 17 and Ber 18 estimate. We believe that this difference arises from the poor fit of the clump stars in the Phelps (1997) study. To cast light on this point, we overimposed (see Fig. 5a) to the cluster CMD a 13 Gyr isochrone. Fixing the clump color and magnitude, the TO appears too faint, and the RGB too red. On the other hand, trying to fix the color and magnitude of the TO, which is more blurred, the RGB clump turns out to be bluer and brighter than observed. This is confirmed by Carraro et al. (1999), who find that Berkeley 17 is about 9 Gyr old analyzing published BV I photometry The luminosity function of Berkeley 17 Fig. 5a and b. Comparison of clusters CMD with theoretical isochrones. Panel a shows the CMD of Berkeley 17, with overimposed two Z = isochrones for the ages of 9 Gyr (solid line) and 13 Gyr (dashed line). Reddening and distance modulus are 0.30 mag and mag, respectively. Panel b shows the CMD of Berkeley 18, with overimposed a Z = isochrone of 4 Gyr. Reddening and distance modulus are 0.25 mag and mag, respectively. hand, Phelps (1997) reached considerably different results, although the two photometric studied were not significantly different (see Carraro et al. 1999). In details he found an age of Gyr, a distance modulus (m M) 13.9 mag and a reddening in the range 0.52 < E(B V ) < 0.68 mag and 0.61 < E(V I) < 0.71 mag. Adopting the relation (Girardi et al. 1998): [F e/h] =log Z (4) we can derive a theoretical metal abundance Z This value is adopted to generate isochrones from the theoretical models by Girardi et al. (1999). The best fit is shown in Fig. 5a. It has been obtained overimposing an 9 Gyr isochrone shifted by (m M) K = and E(J K) = These value are uncertain within 0.1 mag. From the relation E(J K) =0.52 (5) E(B V ) taken from Cardelli et al. (1989), E(J K) transforms into E(B V )=0.58 mag. As a consequence, (m M) o becomes This puts Berkeley kpc away from the Sun. Reddening and distance are in agreement with Phelps (1997) results. As for the age, we find a value consistent with Kaluzny (1994) suggestions, but significantly lower than Phelps (1997) To confirm the age determination based on the CMD, we compare the luminosity function (LF) of the MS of the cluster (defined as stars fainter than K =15or bluer than (J K) =0.75) with the simulations of two populations of age 8 Gyr and 13 Gyr having Z = First, by means of the usual experiments with artificial stars we derive the incompleteness correction γ for Be 17, defined as the ratio of recovered on added stars. It turns out to be γ = 1, 0.86, 0.20 for K<15.5, 16, 17 mag. The theoretical LFs are corrected for incompleteness due to crowding, dividing them by γ. Second, we derive the LF of the disc stars contaminating the cluster using the model of the Galactic disc proposed by by Vallenari et al This model has been tested fitting CMD and luminosity functions of low latitude Galactic fields (see Schmidtobreick et al for details). Since this cluster is located at low Galactic latitude, the main contamination is given by the thin disc. Starting from a double exponential density distribution of the Galactic thin disc, assuming a scale height of 250 pc, a scale length of 2000 pc, constant star formation rate from 10 Gyr to 0.1 Gyr, we derive the simulated LF of the disc component, imposing the number of stars observed in the range 12 <K<14. Changing the input parameters of the simulations (scale height and scale length of the disc to 300 pc and 3000 pc, respectively), we estimate that the error on the disc LF at the level of the expected turnoff of Berkeley 17 (i.e. between 14 and 15 mag) is about 10 15%. Finally the observational LF of the MS is obtained, subtracting the disc MS contamination by the initial LF. In Fig. 6 we compare the simulated LFs with the data. For magnitudes fainter than K 16 the incompleteness correction becomes relevant and the comparison is not significant. For brighter magnitudes, the oldest age gives a turnoff fainter than the observed one, whereas the youngest one is in reasonable agreement with the data. This result might be weakened by the fact that the LF of the disc population has not been derived from observations, but simulated. However, to reproduce the data with an age of 13 Gyr, we need to increase the disc population of about 30%, which is much higher than the estimated uncertainty. As a conclusion, on the basis of the analysis of the CMD and LF, although an age of 13 Gyr cannot be completely ruled out, a much younger age ( 8 Gyr) seems to be favored.

5 G. Carraro et al.: Near IR photometry of the old open clusters Ber 17 and Ber Table 3. Derived parameters of the studied clusters. Cluster [F e/h] Age E(B V ) (m M) o Distance dex Gyr mag mag kpc Berkeley ± Berkeley ± an apparent distance modulus (m M) K = mag.the errors in these estimates are around 0.1 mag. The corresponding visual reddening E(B V ) turns out to be 0.40 mag from Eq. (5), in agreement with Kaluzny s (1997) suggestions. Using this value, we estimate the distance of Berkeley 18 to the Sun to be about 4.5 kpc, significantly lower than Kaluzny s (1997) estimate. 6. Conclusions We have presented and discussed near IR photometry of two poorly studied old open clusters, Berkeley 17 and Berkeley 18. From the analysis of the CMDs we derived cluster fundamental parameters which are listed in Table 3. In details, we derived a photometric estimate of the cluster metal abundance (Column 2), new age estimates with uncertainties derived from the fit of several isochrones of different ages (Column 3), reddening (Column 4), distance modulus (Column 5) and heliocentric distance (last column). The main results of this study can be summarized as follows. Fig. 6. The LF of MS of Berkeley 17 is compared with theoretical simulations with ages of 8 and 13 Gyr corrected for incompleteness. The disc population is simulated and subtracted from the observed LF. The long dashed line represents the original data, dashed-dotted the disc population, solid the MS LF of Berkeley 17 decontaminated by disc stars, and finally short-dashed the simulation of the cluster for the labelled age, 5. Berkeley 18: cluster parameters 5.1. Metallicity No spectroscopic determinations of Berkeley 18 metallicity are available, so we rely only on the slope of the RBG, and derive metallicity from Eq. (3). We obtain (J K)/ K = ± 0.05, which translates in solar metal abundance [F e/h] Age, distance and reddening Berkeley 18 has been studied by Kaluzny (1997). From this preliminary work, Kaluzny suggested that Berkeley 18 is as old asm67(4 Gyr, Carraro et al. 1996), its reddening E(B V ) is probably higher than 0.46 mag, and the distance from the Sun is around 5.8 kpc. In Fig. 5b we present the CMD of Berkeley 18 with a best fitting isochrone of 4 Gyr for the metal abundance Z = From this fit we derive a reddening E(J K) =0.25 mag and using the calibration between the slope of the RGB and the metallicity (Tiede et al. 1997), we estimate that Berkeley 17 has a metal abundance [F e/h] 0.35, while Berkeley 18 has about solar metal abundance; Berkeley 17 is found to be 9 Gyr old. The difference with Phelps (1997) age estimate (about 12 Gyr)is probably due to a better fitting of the clump stars, which are more clearly visible in our study. This younger age is also suggested by the analysis of the cluster LF corrected for the disc star contamination. This new age determination implies that Berkeley 17 is no longer the oldest known open cluster, and suggests the possibility that the Milky Way experienced a minimum of star formation between 10 and 13 Gyr ago; Berkeley 18 is as old as M 67, but its distance is significantly lower than previous estimates. Acknowledgements. We thank L. Hunt and F. Mannucci for kind assistance during data reduction. This study has been financed by the Italian Ministry of University, Scientific Research and Technology (MURST) and the Italian Space Agency (ASI). The work of Léo Girardi is funded by the Alexander von Humboldt-Stiftung. References Cardelli J.A., Clayton J.C., Mathis J.S., 1989, ApJ 345, 245 Carraro G., Girardi L., Bressan A., Chiosi C., 1996, A&A 305, 849 Carraro G., Girardi L., Chiosi C., 1999, MNRAS, submitted

6 830 G. Carraro et al.: Near IR photometry of the old open clusters Ber 17 and Ber 18 Eggen O.J., Linden-Bell D., Sandage A.R., 1962, ApJ 136, 748 Friel E.D., Janes K.A., Hong L., Lotz J., Tavarez M., 1995, In: Alfaro E.J., Delgado A.J. (eds.) The Formation of the Milky Way. Cambridge University Press, p. 189 Girardi L., Bressan A., Bertelli G., Chiosi C., 1999, in preparation Gratton R.G., Fusi Pecci F., Carretta G., et al., 1997, ApJ 491, 749 Hunt L., Testi L., Borelli S., Maiolino R., Moriondo G., 1994, Arcetry Observatory Technical Report N.4/94 Hunt L.K., Mannucci F., Testi L., et al., 1998, AJ 115, 2594 Janes K.A., Phelps R.L., 1994, AJ 108, 1773 Kaluzny J., 1994, Acta Astron. 44, 247 Kaluzny J., 1997, A&AS 121, 455 Phelps R.L., 1997, ApJ 483, 826 Schmidtobreick L., Vallenari A., Bertelli G., 1998, AGM 14, 63 Searle L., Zinn R., 1978, ApJ 225, 357 Setteducati A.F., Waever M.F., 1962, In: Newly Found Stellar Clusters. Radio Lab. Univ. of California, Berkeley Tiede G.P., Martini P., Frogel J.A., 1997, AJ 114, 694 Vallenari A., Bertelli G., Schmidtobreck L., 1999, A&A, submitted Vallenari A., Carraro G., Girardi L., Richichi A., 1999, A&A, in preparation